Agricultural and Biosystems Engineering
Conference Proceedings and Presentations
Agricultural and Biosystems Engineering
7-2005
Evaluation of Phosphate Ion-Selective Membranes
for Real-time Soil Nutrient Sensing
Hak-Jin Kim
University of Missouri
John W. Hummel
United States Department of Agriculture
Stuart J. Birrell
Iowa State University, [email protected]
Kenneth A. Sudduth
United States Department of Agriculture
Follow this and additional works at: http://lib.dr.iastate.edu/abe_eng_conf
Part of the Agriculture Commons, Bioresource and Agricultural Engineering Commons, and the
Soil Science Commons
The complete bibliographic information for this item can be found at http://lib.dr.iastate.edu/
abe_eng_conf/409. For information on how to cite this item, please visit http://lib.dr.iastate.edu/
howtocite.html.
This Conference Proceeding is brought to you for free and open access by the Agricultural and Biosystems Engineering at Iowa State University Digital
Repository. It has been accepted for inclusion in Agricultural and Biosystems Engineering Conference Proceedings and Presentations by an authorized
administrator of Iowa State University Digital Repository. For more information, please contact [email protected].
An ASAE Meeting Presentation
Paper Number: 051033
Evaluation of Phosphate Ion-Selective Membranes for
Real-time Soil Nutrient Sensing
Hak-Jin Kim, Graduate Research Assistant
Biological Engineering Department, University of Missouri, Columbia, Missouri 65211, USA
[email protected]
John W. Hummel, Agricultural Engineer
Cropping Systems & Water Quality Research Unit, USDA-ARS, Columbia, Missouri 65211,
USA, [email protected]
Stuart J. Birrell, Associate Professor
Department of Agricultural and Biosystems Engineering, Iowa State University, Ames, Iowa
50011-3080, USA, [email protected]
Kenneth A. Sudduth, Agricultural Engineer
Cropping Systems & Water Quality Research Unit, USDA-ARS, Columbia, Missouri 65211,
USA, [email protected]
Written for presentation at the
2005 ASAE Annual International Meeting
Sponsored by ASAE
Tampa Convention Center
Tampa, Florida
17 - 20 July 2005
Abstract. A real-time soil nutrient sensor would allow the efficient collection of data with a fine spatial
resolution, to accurately characterize within-field variability for site-specific nutrient application. Our
goal was to evaluate the applicability of a phosphate membrane to the measurement of phosphate
levels in soil extractants and to determine how previously developed nitrate and potassium
The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the
official position of the American Society of Agricultural Engineers (ASAE), and its printing and distribution does not constitute an
endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASAE
editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an
ASAE meeting paper. EXAMPLE: Author's Last Name, Initials. 2005. Title of Presentation. ASAE Paper No. 05xxxx. St. Joseph, Mich.:
ASAE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASAE at [email protected] or
269-429-0300 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).
membranes would be affected by the presence of phosphate. A type of PVC-based phosphate
membrane containing an organotin compound, bis(p-chlorobenzyl)tin dichloride, was evaluated,
along with the nitrate and potassium membranes, in pH 7 Tris buffer solution and Kelowna soil
extractant for sensitivity and long-term stability. The phosphate membranes in the Tris buffer solution
of pH 7 exhibited a response over a range of 10-5 to 10-1 mol/L phosphate concentrations with an
average slope of -28.2 +1.5 mV per activity decade of dibasic phosphate. The response speed of
tested electrodes containing phosphate, nitrate and potassium membranes was rapid, reaching an
equilibrium response in less than 15 s. However, the phosphate membrane in the Kelowna solution
of pH 8.5 was almost insensitive to different phosphate levels from 10-6 to 10-2 mol/L due to the
presence of a high concentration of fluoride in the solution. In addition, the tin compound-based
phosphate membranes had limited lifetimes of less than 14 days. It is not expected that the tested
phosphate membranes could be used for phosphate detection in other soil extractants, such as Bray
P1 and Mehlich III solutions, because they also contain high concentrations of fluoride.
Keywords. ion-selective membranes, ion-selective electrode (ISE), soil testing, phosphate, soil
extractants, Kelowna, sensitivity, response time
The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the
official position of the American Society of Agricultural Engineers (ASAE), and its printing and distribution does not constitute an
endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASAE
editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an
ASAE meeting paper. EXAMPLE: Author's Last Name, Initials. 2005. Title of Presentation. ASAE Paper No. 05xxxx. St. Joseph, Mich.:
ASAE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASAE at [email protected] or
269-429-0300 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).
Introduction
The soil macronutrients, nitrogen (N), phosphorus (P), and potassium (K), are the most
important nutrients for crop growth. These nutrients in the soil solution are taken into plants in
various ionic forms such as nitrate (NO3-), orthophosphates (H2PO4- or HPO42-), and potassium
(K+) through a combination of root interception, mass flow and diffusion processes. Accurate
measurement of soil macronutrients can be used to more accurately estimate needed fertilizer
application rates, increasing the efficiency of soil fertility management, and improving the
profitability of crop production.
Excessive use of commercial fertilizer has been cited as a source of contamination of surface
and groundwater (Staver and Brinsfield, 1990; Mallarino, 1998). High levels of phosphorus in
the soil may leach into water ecosystems and create an imbalance, resulting in excessive
growth of algae in lakes and rivers. Optimum application rates can reduce the potential for
environmental pollution created by excessive application of chemical fertilizers. Site-specific
sensing of macronutrients would make it possible to characterize within-field variability, and lead
to fertilizer application rates that are optimized for each sub-field area.
Monitoring phosphorus in soils is typically performed by compositing soil from a number of
sampling sites into one sample for analysis using conventional soil laboratory testing methods.
Various analytical methods have been used for the determination of phosphorus, mostly based
on spectrophotometric techniques for detecting a colored complex formed by the reaction of
phosphorus with a molybdate ion (Brown, 1998). However, sample collection and analysis
methods are inherently costly and time consuming, thereby limiting the number of samples
tested.
Recently, the majority of the research on methods for the determination of phosphorus has
concentrated on the use of ion-selective electrodes (ISEs), which respond to monobasic
(H2PO4- ) or dibasic (HPO42-) phosphate forms, and flow injection analysis (FIA), because of
advantages over spectrophotometric methods, such as fast response and low cost. Several
researchers reported on the development of phosphate ISEs using PVC-based membranes to
detect phosphates in biological samples (Glazier and Arnold, 1988; 1991; Carey and Riggan,
1994; Liu et al., 1997; Fibbioli et al., 2000; Wroblewski et al., 2001).
Among the PVC-based membranes, a cyclic polyamine ionophore (Carey and Riggan, 1994)
provided good selectivity and favorable sensitivity with a detection limit of 10-5 mol/L dibasic
phosphate in a solution with pH controlled at 7.2. Also, a tin compound ISE containing bis(pchlorobenzyltin) dichloride as the ionophore was developed by Glazier and Arnold (1988 and
1991), and showed good selectivity for dibasic phosphate over nitrate and chloride.
Cobalt wire ISEs were useful in detecting monobasic phosphate in a potassium
hydrogenphthalate solution of pH 5 (Chen et al., 1997; Marco et al., 1998). The solid-state ionselective electrodes using cobalt wires have been used to monitor phosphates in waste waters
and fertilizers (Xiao et al., 1995; Meruva and Meyerhoff, 1996; Chen et al., 1997; Engblom,
1998; Marco et al., 1998). In particular, Engblom (1998) studied the applicability of a metallic
cobalt wire electrode to the measurement of phosphate in ammonium lactate-acetic acid (AL)
extracts commonly used in Sweden. As a result, the cobalt electrodes were applicable to
phosphate detection in soil extracts. However, the sensor was significantly affected by organic
substances and pH in soil extracts, thereby resulting in reduced sensitivity.
Despite this continuing effort, few phosphate sensors have become commercially available,
because the design of a carrier for selective recognition of orthophosphates is especially
challenging (Tsagatakis et al., 1994; Fibbioli et al., 2000). According to Tsagatakis et al. (1994),
2
the free energy of the phosphate species is very small and the large size of orthophosphate
prohibits the use of size-exclusion principles for increased selectivity. Another limitation is that
the response of phosphate ISEs is dependent on the solution pH, since the ionic forms of
phosphate in the solution vary as pH changes.
As we have previously evaluated nitrate and potassium membranes (Kim et al., 2003; 2004),
the aim of this study was to investigate the applicability of a PVC-based phosphate membrane
containing an organotin compound (Glazier and Arnold, 1988; 1991) to the determination of
phosphates in soil extractants. For this purpose, the response characteristics of the phosphate
membranes were evaluated, along with the nitrate and potassium membranes developed in
previous studies, in terms of sensitivity and repeatability. In addition, the effects of base solution
and membrane age on sensing performance were investigated.
Materials and Methods
Reagents
PVC-based phosphate ion-selective membranes were prepared using a tin compound, bis(pchlorobenzyl)tin dichloride. The phosphate ionophore was synthesized according to the
procedures outlined in Glazier (1988). Dibutyl sebacate as a plasticizer, N, Ndimethylformamide as a solvent for organic compounds, polyvinyl chloride (PVC), and
tetrahydrofuran (THF) were purchased from Sigma-Aldrich Corp. (St. Louis, Mo.).
One nitrate ion-selective membrane reported in previous papers (Kim et al, 2003; 2004), was
prepared using a quaternary ammonium compound, tetradodecylammonium nitrate (TDDA),
and a plasticizer, nitrophenyl octyl ether (NPOE). Valinomycin, bis(2-ethylhexyl) sebacate
(DOS) and potassium tetrakis (4-chlorophenyl) borate (KTpClPB) were used for the preparation
of potassium ion-selective membranes. These chemicals also were purchased from SigmaAldrich Corp.
Two different base solutions, pH 7 Tris buffer solution and Kelowna soil extracting solution, were
prepared using distilled and deionized water with a specific resistance of 18.0 MΩ cm-1
produced by a distilled water system (Model MP-6A, Corning). The pH 7 buffer solution
consisted of 0.01 mol/L tris(hydroxymethyl) aminomethane (Tris, Fisher Scientific) with 0.0045
mol/L H2SO4 (sulfuric acid, Sigma-Aldrich) and the Kelowna extractant solution contained 0.25
mol/L CH3COOH (acetic acid, Fisher Scientific) and 0.015 mol/L NH4F(ammonium fluoride,
Sigma-Aldrich).
All other chemicals used -- potassium dibasic phosphate (K2HPO4), sodium nitrate (NaNO3),
lithium acetate (LiAc), potassium chloride (KCl), and ammonium hydroxide (NH4OH), were of
analytical reagent grade and purchased from Sigma-Aldrich Corp. (St. Louis, Mo.) and Fisher
Scientific (Cincinnati, Ohio).
Preparation of Ion-Selective Membranes and Electrodes
The phosphate membrane-casting solution was prepared as reported in previous studies
(Glazier and Arnold, 1991), and contained 70.2 mg (18% wt) of bis(p-chlorobenzyl)tin dichloride,
133.5 mg (34% wt) of PVC, 141.9 mg (36% wt) of dibutyl sebacate, and 48.3 mg (12% wt) of N,
N-dimethylformamide in 3 mL of THF. Phosphate membranes were formed, as previously
described by Glazier and Arnold (1988; 1991), by dipping the free ends of Hitachi ISE electrode
bodies in the casting solution three times. Membranes were allowed to dry after the first two
3
dips and then they were stored overnight following the final dip. Phosphate ISEs were
constructed by using 0.1mol/L KCl as an internal filling solution in the electrode body and
inserting an Ag/AgCl reference electrode into the top. The electrodes were conditioned
overnight in the Tris buffer solution. Prior to testing, the electrodes were immersed in 0.01M
phosphate solution three times for a few minutes so that steady electrical potentials could be
obtained in the presence of phosphate.
Ion-selective membranes for nitrate and potassium were prepared based on TDDA and
valinomycin ionophores as reported in previous papers (Kim et al., 2003; 2004). The nitrate ionselective membrane was obtained by casting a mixture of TDDA (30 mg, 15% wt), NPOE (80
mg, 40% wt), and PVC (90 mg, 45% wt) in 2 mL of THF. The composition of the potassium ionselective membrane prepared was 4 mg (2% wt) of valinomycin, 1 mg (0.5% wt) of lipophilic
additive (KTpClPB), 129.4 mg (64.70% wt) of DOS plasticizer, and 65.6 mg (32.80% wt) of PVC
in 2 mL of THF. The casting solutions for both nitrate and potassium membranes were poured
into 23-mm glass rings resting on polished glass plates, and allowed to evaporate for 24 h at
room temperature. The membranes, formed as a film, were removed from the glass plate.
Membrane disks, cut with a diameter of 2.5 mm from the membrane, were attached to the ends
of the Hitachi ISE electrode bodies using THF solution. Each nitrate ISE electrode was filled
with an internal solution consisting of 0.01mol/L NaNO3 and 0.01mol/L NaCl. Potassium chloride
(0.01mol/L) was employed as the internal reference solution of the potassium ISE electrodes.
The nitrate and potassium ISE electrodes were stored overnight separately in 0.01 mol/L
NaNO3 and 0.01 mol/L KCl solution, respectively.
A double junction Ag/AgCl electrode (Model PHE 3211, Omega Engineering, Stamford, Conn.)
was used as the reference electrode. To dissuade contamination of sample analyte ions such
as K+ and NO3- by the reference electrode, 1mol/L LiAc was used as the outer reference
solution in the reference electrode.
EMF Measurements
A test apparatus (fig. 1) was used for automatically controlling the system based on userdefined parameters and simultaneously recording EMF (electromotive force) values of 16
electrodes. Details of this equipment were described previously (Kim et al., 2003; 2004).
Calibration Solution
Controller
Valve
Pump
Data Acquisition System
ISE
Computer
Ref. Electrode
Motor
Figure 1. Schematic representation of the automated test apparatus.
4
Data collection was conducted at 15 s and 60 s after the injection of each test solution into the
test stand. At each of the two data collection times, three measurements, each consisting of the
mean of a 0.1-s burst of 1 kHz data, were obtained on a 3-s interval and averaged.
Sensitivity Tests
The response characteristics of the electrodes were examined by measuring the EMFs of each
ISE in six standard solutions of K2HPO4 containing from 10-6 to 10-1 mol/L concentrations. The
standard solutions were prepared by successive 10:1 dilutions of the 0.1 mol/L concentration
using each of two different base solutions (the Tris buffer and the Kelowna solution).
Each test included phosphate, nitrate, and potassium ISEs. A set of test phosphate ISEs
included four different ages, i.e., time between membrane preparation and test: six electrodes of
age 4 days, and one each of 14, 20, and 33 days, in order to determine how the responses of
the ISEs to phosphate changed as the membranes aged. The electrodes with potassium and
nitrate membranes were also tested to investigate how those membranes would be affected by
the presence of phosphate and potassium. As a result, nine phosphate ISEs, two potassium
ISEs, and two nitrate ISEs were included in the test set.
Since the phosphate species in solution is a function of pH (Lindsay, 1979), fig. 2a), the pHs of
all tested Tris buffer standards containing different phosphate concentrations were adjusted to
7.00+0.01, as measured with a combination pH electrode (Model 81-72, Orion, Cambridge,
Mass.) and a pH meter (Model SA-720, Orion, Cambridge, Mass.), through the addition of
sulfuric acid. Duplicating this pH level used by Glazier and Arnold (1991) allowed a comparison
with those results, even though, at this pH level, a portion of the phosphate is not in the dibasic
form detected by the ISE. When using the Kelowna extractant as the base solution, the pHs of
the standard solutions were readjusted to 8.5+0.01, where the predominant form is dibasic
phosphate. Another advantage was that pH 8.5 was above the range of pH where small
additions of a base solution produce rapid pH changes (fig. 2b).
The sensitivity tests were repeated three times. The EMF values of the all ISEs at different
K2HPO4 concentrations (10-6, 10-5, 10-4, 10-3, 10-2, and 10-1 mol/L K2HPO4) were determined in
each test sequence with the automated test stand.
(a)
(b)
10
8
pH
6
4
2
0
0
10
20
30
40
50
1M NH4OH (mL)
Figure 2. Distribution of orthophosphate ions depending on pH level (a, Lindsay (1979))
and titration curve for 90mL of Kelowna solution (b).
5
Since standard potentials among electrodes normally vary due to difference in internal
resistance and thickness of the membrane (Carey and Riggan, 1994), the electric potential was
normalized by setting the EMF values obtained at 10-6M phosphate concentration in the third
replication to 0 mV. This procedure removed variability between electrodes in terms of standard
potential, while allowing differences between replications to be evaluated.
To calculate sensitivity slopes for phosphate ISEs in tested concentration ranges, each activity
of dibasic and monobasic phosphate species in solution was calculated using an iterative
method. The approach considers change in ionic strength and uses known equilibrium
constants for the reaction of phosphates in solution, because the ionic strength is a function of
the solution pH due to phosphate species equilibria with the hydrogen ion activity (Lindsay,
1979; Carey and Riggan, 1994).
The total phosphate concentration in the pH range of 4~10 can be calculated:
[ PO 4 ] total = [ H 2 PO 4− ] + [ HPO 42 − ]
(1)
where [PO4] is total phosphate concentration, and [H2PO4] and [HPO4] are concentrations of
monobasic and dibasic phosphates, respectively.
The equilibrium constant between monobasic and dibasic phosphates can be represented:
log
[ H 2 PO4− ]
= 7.20 − pH
[ HPO42− ]
(2)
The ionic strength was calculated using the estimated concentrations, and the activity
coefficients for the dibasic phosphate species were then estimated using the Debye-Hückel
formula (Lindsay, 1979; Eggins, 2002).
Results and Discussion
Response Characteristics in Tris Buffer and Kelowna Solutions
The response (EMF) curves of the six newest phosphate ISEs (4 days old at the time of testing),
two nitrate and two potassium ISEs to different potassium phosphate (K2HPO4) concentrations
ranging from 10-6 mol/L to 10-1 mol/L in pH 7 Tris buffer and Kelowna solutions are shown in
figure 3. In each of the three replicates of the test sequence, successively more concentrated
test solutions were presented to the ISEs. It is evident that the phosphate electrodes in the Tris
buffer solution (fig. 3a) were sensitive to different phosphate concentrations and the responses
were repeatable during three replicate measurements. Similarly, the potassium ISEs responded
to potassium with consistent sensitivity (fig. 3b). The nitrate ISEs had a slight sensitivity to
phosphate (fig. 3b) with a decrease in EMF (<15 mV) at 10-1 mol/L phosphate concentration.
The use of Kelowna solution influenced the responses of all ISEs significantly. In particular, as
shown in figure 3c, the responses of the phosphate ISEs in the Kelowna solution were
decreased considerably, thereby resulting in little change in EMF in the range of 10-6 to 10-2
mol/L phosphate concentration. Similarly, at low potassium concentrations below 10-3 mol/L,
there appeared to be little change in response for the potassium membranes (fig. 3d). However,
the potassium ISEs exhibited a linear response over a range of 10-3 to 10-1 mol/L potassium
concentrations.
6
200
(a)
150
(1)
300
P-03
P-06
P-02
P-05
100
EMF (mV)
(5)
(6)
0
0
-100
-50
-200
-300
-100
0
500
1000
1500
2000
Time (s)
2500
3000
0
3500
300
(c)
P-01
P-04
50
P-03
P-06
P-02
P-05
500
1000
(d)
1500
2000
Time (s)
2500
N-01
K-01
200
3000
3500
N-02
K-02
100
EMF (mV)
0
EMF (mV)
N-02
K-02
(3)
50
100
N-01
K-01
(b)
200
(2)
(4)
100
EMF (mV)
P-01
P-04
-50
0
-100
-100
-150
-200
-300
-200
0
500
1000
1500
2000
Time (s)
2500
3000
3500
0
500
1000
1500
2000
Time (s)
2500
3000
3500
Figure 3. Response test profiles for different K2HPO4 concentrations: (a) phosphate membrane response in Tris buffer, (b)
nitrate and potassium membrane response in Tris buffer, (c) phosphate membrane response in Kelowna extracting
solution, and (d) nitrate and potassium membrane response in Kelowna extracting solution. The numbers in (a) identify the
different K2HPO4 concentrations: (1) 10-6; (2) 10-5; (3) 10-4; (4) 10-3; (5) 10-2; and (6) 10-1 mol/L.
7
60
40
300
(a)
250
Measure EMF (mV)
Measure EMF (mV)
20
0
-20
P-01
P-02
P-03
P-04
P-05
P-06
-40
-60
-80
200
150
K -01
K -02
N-01
N-02
100
50
0
-100
-120
-120 -100
(b)
-50
-80
-60 -40 -20
0
Premeasure EMF (mV)
20
40
60
-50
0
50
100
150
200
Premeasure EMF (mV)
250
300
Figure 4. Relationship between EMF values measured 15s and 60s after the injection of
test solutions for phosphate (a) and nitrate and potassium membranes (b).
A study of the response speed of each membrane type was conducted by relating the EMF
values taken at 15 s (premeasure) to those obtained at 60 s (measure) after each test solution
was introduced. As shown in fig. 4, the measure EMF (Y) values were highly correlated with the
premeasure EMFs (X), showing an almost 1:1 relationship between the two values: Y= 0.96X 0.41 (R2=0.99**) for phosphate and Y= 0.99X + 3.39 (R2=0.99**) for nitrate and potassium ISEs.
Therefore, it was evident that the ISEs could reach an equilibrium response in less than 15 s.
Variability of response between membranes
-40
(a)
20
P-01
P-02
P-03
P-04
P-05
P-06
15
10
5
0
10-7
Slope (mV / activity decade of H2PO )
25
10-1 to 10-5 M range
10-1 to 10-4 M range
(b)
4
Variability in EMF between replications (mV)
The variability of response among the six tested phosphate ISEs was examined by comparing
the standard deviations in EMF measured with the ISEs for three replicate measurements of
phosphates and average sensitivity of each ISE (fig. 5). One electrode, P-03, showed relatively
poor repeatability (fig. 5a) with a standard deviation in EMF of >10 mV. Comparing sensitivity
slopes in the concentration ranges of 10-5 to 10-1, and 10-4 to 10-1 mol/L phosphate (fig. 5b), one
electrode, P-04, showed less sensitivity than did the other electrodes. Obviously, these two
electrodes (P-03 and P-04) were producing questionable data. Based on data obtained with the
other four electrodes (P-01, P-02, P-05, and P-06), phosphate ISEs responded to phosphate
10-6
10-5
10-4
10-3
10-2
10-1
100
-30
-20
-10
0
P-01
P-02
P-03
P-04
P-05
P-06
K2HPO4 concentration (mol/L)
Figure 5. Comparison of phosphate ISEs in terms of standard deviation of EMF values
(a) and sensitivity slope (b).
8
over a range of 10-5 to 10-1 mol/L with an average slope of -28.2 mV per activity decade of
HPO42-, yielding a standard deviation in EMFs of 5.3 + 3.0 mV for three replicate
measurements.
Sensitivity of membranes in Tris buffer and Kelowna solutions
Sensitivity of each membrane type to varying phosphate concentrations was calculated when
using the Tris buffer and the Kelowna solution as base solutions (fig. 6). In general, in the Tris
buffer solution (pH = 7.00+0.01), the EMF values obtained with the phosphate membranes (fig.
6a) were almost linearly proportional to the logarithm of phosphate concentration in the range of
10-4 to 10-1 mol/L with a sensitivity slope of -33.1 + 1.5 mV per activity decade of HPO42-, which
is comparable to the sensitivities reported in previous studies (Glazier and Arnold, 1988). In
contrast, in the Kelowna solution (pH = 8.5 + 0.01), the four phosphate membranes were almost
insensitive to phosphate (fig. 6c), regardless of the level of phosphate in the tested solutions.
The potassium membranes in the Tris buffer solution (fig. 6b) showed a slope of 50.3 + 1.3 mV
per activity decade of K+. In the Kelowna solution (fig. 6d), at low potassium concentrations
below 10-3 mol/L, the sensitivity of potassium membranes was considerably decreased, thereby
40
(a)
20
250
-40
-60
150
100
-80
50
-100
0
-120
10-7
40
10-6
(c)
20
200
P-01
P-02
P-05
P-06
-20
-40
-60
N-01
N-02
K-01
K-02
10-6
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
(d)
100
50
0
N- 01
N- 02
K -01
K -02
-100
-100
10-6
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
100
100
150
-50
-80
-120
10-7
-50
10-7
100
EMF (mV)
EMF (mV)
0
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
(b)
200
-20
EMF (mV)
EMF (mV)
0
300
P-01
P-02
P-05
P-06
-150
10-7
10-6
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
100
Figure 6. Response of each membrane to different K2HPO4 concentrations: (a)
phosphate membrane response in Tris buffer, (b) nitrate and potassium membrane
response in Tris buffer, (c) phosphate membrane response in Kelowna extracting
solution, and (d) nitrate and potassium membrane response in Kelowna extracting
solution.
9
resulting in a detection limit of about 10-3 mol/L, which is higher than that (10-4 mol/L) obtained in
previous studies (Kim et al., 2004). Such a decrease in sensitivity for the potassium
membranes, as compared to that in previous tests, occurred because of the presence of a high
concentration (about 0.2 mol/L) of ammonium (NH4+), which was introduced when ammonium
hydroxide was added to adjust the pH of the Kelowna solution.
Effects of base solution type and membrane age on sensitivity
As observed from a plot (fig. 7a) comparing responses of the phosphate membranes in different
base solutions, the average EMF value of the phosphate ISEs in the Tris buffer solution
decreased by about 100 mV as the phosphate concentration increased from 10-6 mol/L to 10-1
mol/L, whereas the decrease over the same concentration range obtained in the Kelowna
solution was only about 13 ~18 mV.
Such a significant decrease in sensitivity for the phosphate membranes may be associated with
the presence of a high concentration of fluoride (0.015 mol/L) in the Kelowna solution. Previous
studies by Glazier and Arnold (1991) show that the selectivity coefficient of the membrane for
fluoride is 0.279, which means that the tin compound phosphate membrane is only about 3.58
times more sensitive to dibasic phosphate than to fluoride. When fluoride and dibasic phosphate
having the same concentration are dissolved in solution, the ionic activities for fluoride are larger
than those for dibasic phosphate, since there is a greater decrease in ionic activity for dibasic
phosphate than for fluoride. For example, at 0.1 mol/L total phosphate concentration, the ionic
activity of dibasic phosphate in the pH 8.5 Kelowna solution was approximately 0.01, which is
nearly the same as that of 0.015 mol/L fluoride concentration. This means the sensitivity in the
0.1 mol/L phosphate standard may be reduced by about 8 mV (27.9% of 28.2 mV/decade in a
range of 10-5 to 10-1 mol/L) due to interference by the fluoride ion. The reduced sensitivity of
about 20 mV for the phosphate concentration change from 0.01 mol/L to 0.1 mol/L is of similar
magnitude to the sensitivity of -15 ~ -18 mV/decade obtained in this experiment.
The changes in response to phosphate when electrodes having different ages were
simultaneously tested are shown in figure 7b. The electrodes were stored in the pH 7 Tris buffer
at room temperature (22.5 to 23.5 °C) between measurements. As shown in figure 7b, the
responses of the electrodes dramatically deteriorated as the electrodes aged. After 14 days of
use, an increase in detection limit from 10-5 to 10-4 ~10-3 mol/L total phosphate concentration
and a much shorter linear range were observed. Possible causes of the deterioration of
electrode response are rapid leaching of the tin compound ionophore from the membrane or a
rapid breakdown of the tin compound structure.
(a)
20
0
0
-20
-20
EMF (mV)
EMF (mV)
20
-40
-60
Kelowna
Tris buffer
-80
-120
10-7
-40
-60
-80
-100
-100
10-6
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
100
(b)
-120
10-7
33 days
20 days
14 days
4 days
10-6
10-5 10-4 10-3 10-2 10-1
K2HPO4 concentration (mol/L)
100
Figure 7. Effects of base solutions (a) and membrane ages (b) on change in electrode response.
10
Summary and Conclusions
The sensitivity and long-term stability of a PVC-based phosphate membrane containing an
organotin compound as an ionophore were evaluated in pH 7 Tris buffer solution and Kelowna
soil extractant to find out whether the phosphate membranes could be of use in the
determination of phosphate in soil extractants. PVC-based nitrate and potassium membranes
developed in previous studies (Kim et al, 2003; 2004) were also included in the test set to
investigate the sensitivity of the nitrate and potassium membranes to phosphate and potassium.
The PVC-based phosphate membrane containing an organotin compound exhibited a linear
response over a range of 10-4 to 10-1 mol/L phosphate concentrations in the Tris buffer of pH 7
with an average slope of -33.1 +1.5 mV per activity decade of dibasic phosphate, which is
comparable to results obtained in previous studies (Glazier and Arnold, 1988; 1991), The
response speed of tested electrodes containing phosphate, nitrate and potassium membranes
was rapid enough to reach an equilibrium response in less than 15 s. The tested potassium
membranes responded to potassium with consistent sensitivity and good repeatability,
consistent with results of previous tests (Kim et al., 2004). Nitrate membranes showed little
response, as expected, to phosphate even though there was a slight decrease in EMF at 10-1
mol/L phosphate concentration in the Tris buffer solution.
The phosphate membrane in the Kelowna solution, which had been adjusted to a pH of 8.5, was
almost insensitive to different phosphate levels from 10-6 to 10-2 mol/L due to the presence of a
high concentration of fluoride in the Kelowna solution. Regrettably, it is not expected that the
responses of the tested phosphate membranes would be different in other common phosphorus
soil extractants, such as Bray P1 and Mehlich III solutions, because they also contain high
concentrations of fluoride. Modification of the composition of the extractant solution to reduce
the level of fluoride may be an area for future work. However, this approach does not appear
promising, because fluoride plays a significant role in preventing the re-adsorption of solubilized
P by soil colloids during extraction.
The limited functional lifetime of less than 14 days exhibited by the phosphate membrane was
less than had been expected. Additional research will be needed to determine if modifications to
storage conditions can extend membrane life.
For future work, evaluation of other types of phosphate membranes is planned, with the goal of
identifying a sensitive membrane composition that is minimally affected by the various ions
present in soil extractants.
Acknowledgements
The authors appreciate the assistance of the following USDA-ARS Cropping Systems & Water
Quality Research Unit employees: Scott T. Drummond -- for contributions to the development of
the computer interface and control software, and Kurtis J. Holiman and Robert L. Mahurin -- for
their assistance in fabrication and maintenance of the test stand. We also thank Dr. Timothy E.
Glass of the University of Missouri Chemistry Department for providing the tin compound
ionophore.
References
Brown, J.R. (ed.) 1998. Recommended chemical soil test procedures for the North Central
Region. Columbia, Mo.: Missouri Agricultural Experiment Station.
11
Carey, C.M., and W.B. Riggan. 1994. Cyclic polyamine ionophores for use in a dibasicphosphate-selective electrode. Anal. Chem. 66: 3587-3591.
Chen, Z., R.D. Marco, and P.W. Alexander. 1997. Flow-injection potentiometric detection of
phoshates using a metallic cobalt wire ion-selective electrode. Analytical
Communications 34: 93-95.
Eggins, B.R. 2002. Chemical Sensors and Biosensors. West Sussex, England: John Wiley and
Sons.
Engblom, S.O. 1998. Determination of inorganic phosphate in a soil extract using a cobalt
electrode. Plant and Soil. 206(2): 173-179.
Fibbioli, M., M. Berger, F.P. Schmidtchen, and E. Pretsch. 2000. Polymeric membrane
electrodes for monohydrogen phosphate and sulfate. Anal. Chem. 72(1): 156-160.
Glazier, S.A. 1988. Phosphate ion-selective electrode development. Unpublished PhD diss.,
Iowa City, IA: University of Iowa, Department of Chemistry.
Glazier, S.A., and M.A. Arnold. 1988. Phosphate-selective polymer membrane electrode. Anal.
Chem. 60: 2540-2542.
Glazier, S.A., and M.A. Arnold. 1991. Selectivity of membrane electrodes based on derivatives
of dibenzyltin dichloride. Anal. Chem. 63: 754-759.
Kim, H.J., J.W. Hummel, and S.J. Birrell. 2003. Evaluation of ion-selective membranes for realtime soil nutrient sensing. ASAE Paper No. 031075. St. Joseph, Mich.: ASAE.
Kim, H.J., J.W. Hummel, and S.J. Birrell. 2004. Evaluation of ion-selective membranes for realtime soil macronutrients sensing. ASAE Paper No. 041044. St. Joseph, Mich.: ASAE.
Lindsay, W.L. 1979. Chemical Equilibria in Soils. New York, N.Y.: John Wiley and Sons.
Liu, D., W.C. Chen, R.H. Yang, G.L. Shen, and R.Q. Yu. 1997. Polymeric membrane phosphate
sensitive electrode based on binuclear organotin compound. Anal. Chim. Acta 338: 209214.
Mallarino, A. 1998. Soil phosphorus testing for crop production and environmental purposes.
Proceedings of the Integrated Crop Management Conference, 185-192. Ames, IA: Iowa
State University.
Marco, R.D., B. Pejcic, and Z. Chen. 1998. Flow injection potentiometric determination of
phosphate in waste waters and fertilizers using a cobalt wire ion-selective electrode.
Analyst 123: 1635-1640.
Meruva, R.K., and M.E. Meyerhoff. 1996. Mixed potential response mechanism of cobalt
electrodes toward inorganic phophate. Anal. Chem. 68: 2022-2026.
Staver, K.W., and R.B. Brinsfield. 1990. Patterns of soil nitrate availability in corn production
systems: Implications for reducing groundwater contamination. J. Soil and Water Cons.
45(2): 318-323.
Tsagatakis, J.K., N.A. Chaniotakis, and K. Jurkschat. 1994. Designing a novel phosphateselective carrier. Helv. Chim. Acta 77: 2191-2196.
Wroblewski, W., K. Wojciechowski, A. Dybko, Z. Brzozka, R.J.M. Egberink, B.H.M. SnellinkRuel, and D.N. Reinhoudt. 2001. Durable phophate-selective electrodes based on uranyl
salophenes. Anal. Chim. Acta 432: 79-88.
Xiao, D., H.-Y. Yuan, J. Li, and R.-Q. Yu. 1995. Surface-modified cobalt-based sensor as a
phosphate-sensitive electrode. Anal. Chem. 67: 288-291.
12